convnext_flyswot

This model is a fine-tuned version of facebook/convnext-base-224-22k on the image_folder dataset. It achieves the following results on the evaluation set:

• Loss: 0.1441
• F1: 0.9592

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 666
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1
No log	1.0	52	0.6833	0.7484
No log	2.0	104	0.3666	0.8750
No log	3.0	156	0.2090	0.9321
No log	4.0	208	0.1478	0.9449
No log	5.0	260	0.1002	0.9518
No log	6.0	312	0.1053	0.9506
No log	7.0	364	0.1182	0.9616
No log	8.0	416	0.1102	0.9592
No log	9.0	468	0.1262	0.9616
0.203	10.0	520	0.1286	0.9616
0.203	11.0	572	0.1355	0.9592
0.203	12.0	624	0.1299	0.9592
0.203	13.0	676	0.1154	0.9592
0.203	14.0	728	0.1385	0.9580
0.203	15.0	780	0.1330	0.9592
0.203	16.0	832	0.1390	0.9592
0.203	17.0	884	0.1386	0.9592
0.203	18.0	936	0.1390	0.9592
0.203	19.0	988	0.1409	0.9592
0.0006	20.0	1040	0.1411	0.9592
0.0006	21.0	1092	0.1413	0.9592
0.0006	22.0	1144	0.1415	0.9592
0.0006	23.0	1196	0.1426	0.9592
0.0006	24.0	1248	0.1435	0.9592
0.0006	25.0	1300	0.1438	0.9592
0.0006	26.0	1352	0.1434	0.9592
0.0006	27.0	1404	0.1437	0.9592
0.0006	28.0	1456	0.1441	0.9592
0.0002	29.0	1508	0.1440	0.9592
0.0002	30.0	1560	0.1441	0.9592

Framework versions

• Transformers 4.17.0.dev0
• Pytorch 1.10.0+cu111
• Datasets 1.18.3
• Tokenizers 0.11.6